"Oh, boy."

"I can't take it seriously."

"This is literally insane."

"This is atrocious."

"Jesus Christ."

There's something about being continually exposed to the wildest dreams of others that wears on the soul. A long-shot bid for greatness, for true love, or for wealth can be inspiring. But watching many, many long shots has a way of making a person cynical.

The same is true for start-up pitches, which are, almost by definition, long shots. A good pitch can be intensely emotional. "You almost want to see the founder cry," says David Tisch, a 30-year-old angel investor and the managing director of the New York program of TechStars, a business incubator that is now in five cities. "You want to feel that for the past 10 years, they've only been thinking about this one thing. That's how you know they're legit."

And yet, when you've seen thousands upon thousands of start-up pitches, when you've been sold every idea a dozen times-;Kickstarter for small business! Second Life 2! MySpace for live music!-;you can be forgiven for being dismissive; or, in the case of Tisch, for yelling "Douchey!" three times in quick succession as he sits in an empty office staring at his computer screen.

On this particular morning, Tisch is focusing his disgust on a twentysomething entrepreneur with a bold plan (aren't they all?) to revolutionize (per usual) the cosmetics industry. She comes recommended by the founder of a billion-dollar e-commerce company. Her start-up has actual customers-;it has pulled in roughly $100,000 in its first six months-;but Tisch has concluded that she isn't capable of turning it into a big company. "Look," he says, talking primarily to himself. "We have to take this seriously. It's funded by people we know. It's a real market. But the founder is a flutey New York person."

He fires off an e-mail to an investor in the company-;"I kind of hate her," his note begins. Tisch, who shoots me a conspiratorial smile as he hits Send, does not in fact hate her. The grandson of the media mogul Larry Tisch, he is one of those New York types who can be both crass and charming at the same time. In this instance, he wants to see how the entrepreneur's investors defend her in the face of criticism. If they respond forcefully, Tisch will put his first impression aside and give her a chance. "I'll watch this one," he says.

The date is January 24, one day after applications were due for TechStars, a three-month mentorship program that is part boot camp, part investment fund. Some 1,480 young companies have filled out a questionnaire and recorded two short videos for the chance to compete for just 14 spots. That works out to an acceptance rate of less than 1 percent. "Look to your right; look to your left," Tisch said at a recruiting event in early January, modifying the Harvard Law School warning to first-year students. "Probably none of you will get in here."

And yet here they are, showing up at happy hours, sending weekly e-mail updates about their progress, or even, in the case of one group of co-founders, flying to Boulder, Colorado, to pay an unannounced visit to TechStars's national headquarters. There are teams from San Francisco, Seattle, Rome, Bucharest, and Tokyo. There are Harvard M.B.A.'s, former investment bankers, and even a guy nominated for a 2012 Academy Award (best documentary). Some entrepreneurs have applied for the money-;each company accepted into TechStars will immediately receive $118,000 in start-up capital-;but most of the strong candidates raised at least a few hundred thousand dollars even before they applied. One company raised $2.5 million.

I spent most of January and February following this mass of founders, almost all of them in their 20s and 30s. I watched dozens of candidates make their cases-;and then sweat and stammer as they were peppered with follow-up questions delivered with calculated hostility meant to test their resolve. ("Can you build this company without him?" was one question Tisch would often direct at one member of a two-person founding team.) I attended the meeting in which a coterie of the city's top investors, TechStars's selection committee, made its recommendations. I saw the relieved expressions of the young founders when they were told, Yes, you are, as you always thought, one of the elite. I talked to the losers after they had gotten the bad news.

Of course, I read the applications. Though the most promising candidates tend to cultivate relationships with Tisch and his deputy, Adam Rothenberg, long before they formally apply, there are always a handful of what Tisch calls "random companies that show up on the last day and make you go, 'Wow.' "

For the moment, the top of the slush pile is occupied by a company that for the purposes of this story I will call CrowdFundMe. The founders are engineers based in the Czech Republic, aiming to build a crowdfunding website for people outside the U.S.

"OK," Tisch narrates. "They met eight years ago while working for a website. So they know each other." That's a good thing-;founders who have long relationships generally have higher rates of success than those who met recently. Another good thing: CrowdFundMe has raised $500,000 from angel investors.

The application notes that the company has a prototype, and it suggests that Tisch e-mail the founders for a password so he can access it. This inconvenience is too much. "You need me to e-mail you for access to your site?" Tisch yells at his screen in disbelief. He pauses, then calmly says to no one in particular, "Next."

The evaluation takes 2 minutes and 35 seconds, about a minute more than Tisch spends on a typical company. "At the end of the day, you're trying to look at somebody and ask, 'Are you the type of person I could foresee building a $500 million company?' " Tisch says. "It's pretty clear in 20 seconds whether that person has it in them or not. You don't need to know that much."

There are, as many have observed, two American economies. Whereas huge swaths of the country remain hobbled by scarce credit, depressed home prices, and high levels of unemployment, the technology hubs in Silicon Valley, Boston, and New York City are booming. "If you're where the start-ups are, you'd never think there was a recession," says Yael Hochberg, an economist who teaches classes on venture capital and entrepreneurship at Northwestern University's Kellogg School of Management. To young people who can write software code, the question of the moment is not whether or not they will be able to find jobs. It is, says Hochberg, "Do you want to raise $200,000 to start a company, or do you want to go work at Groupon?"

As the rest of the country worries about the slow pace of economic recovery, the tech world frets about whether there's a new start-up bubble. "The global economic meltdown had very little impact on our universe," says Brad Feld, a Boulder venture capitalist and a co-founder of TechStars. Four years ago, Feld made an angel investment in Zynga, a company that makes online video games for Facebook users. Today, Zynga trades on the Nasdaq and is worth more than $9 billion. Facebook itself is expected to be worth as much as $100 billion when it goes public later this year.

TechStars and its chief competitor, Y Combinator, have become the easiest way into this world. "There's this whole tech scene, and it can feel like you're either in or out-;whether it's raising money or hiring or whatever," says Josh Wais, the 23-year-old founder of Wantworthy, a 2011 TechStars graduate. His company, which allows people to make online shopping lists, raised $1 million from investors in January. "TechStars puts you on the inside. It's like how people who want to work at Goldman Sachs get an M.B.A. from Wharton."

Companies accepted into TechStars New York get free office space, $18,000 in cash, $100,000 in convertible debt, and access to 150 mentors-;the list includes Fog Creek Software founder Joel Spolsky, venture capitalist Fred Wilson, and Foursquare founder Dennis Crowley. At the conclusion of the program, graduates take the stage at New York City's Webster Hall to pitch their companies to hundreds of reporters and investors. In exchange, the program takes a 6 percent stake in each company.

For most entrepreneurs, that's a small price to pay. Of the 126 companies that have completed a TechStars program, only 8 percent have failed. By comparison, the failure rate for most technology start-ups is thought to be as high as 90 percent. The average TechStars company raises $1.1 million from venture capitalists, and some raise much more. "If you take a really talented group of people and give them intense mentoring, you sharply increase the chances that their company will be successful," says Hochberg. "I tell my M.B.A. students to apply for these programs. The model works."

In 2005, a little-known engineer named Paul Graham launched a new kind of angel investment fund. "Start-ups," Graham later explained in an essay, "are undergoing the same transformation that technology does when it becomes cheaper. It's so cheap to start Web start-ups that orders of magnitudes more will be started." Graham's program, Y Combinator, which is based in Mountain View, California, and is backed by Sequoia Capital, has cranked out some 380 companies to date. It has helped create the social-news site Reddit, the video site Justin.tv, and the cloud computing company Heroku, whose owners sold it to Salesforce.com for $250 million after just three years in business. A 2007 Y Combinator graduate, Dropbox, has reportedly raised $250 million in venture capital at a valuation of more than $4 billion.

Though Graham himself argued that start-up factories (or incubators, or accelerators) like Y Combinator would work best in Silicon Valley, with its large base of entrepreneurs and investors, David Cohen, the TechStars CEO, thought the model could be adapted to other places. A soft-spoken engineer who had founded three companies in Boulder, Cohen had been making angel investments since 2005 but was looking for a way to get closer to the companies he was trying to help.

In late 2006, Cohen wrote an e-mail to Graham asking if he could partner with Y Combinator to create a Boulder branch. Graham, who declined to comment for this story, refused. "He basically told me to fuck off and die," Cohen says, his voice dripping with bitterness. Cohen went ahead with the program anyway, under a new name. He printed a brochure and finagled a meeting with Feld, who in a matter of minutes agreed to invest. "My thinking was that, worst-case scenario, we'd attract 20 new smart entrepreneurs to Boulder," Feld says.

Although TechStars hasn't yet produced the kind of hits that have come out of Y Combinator, it has quietly established itself as a strong second choice, especially for companies outside Silicon Valley. Socialthing, a graduate of the first TechStars class, in 2007, was sold for roughly $10 million, and SendGrid, a 2009 graduate, now employs 65 people and has raised $27 million in venture capital. Its founder, Isaac Saldana, a Mexican American engineer from East Los Angeles, applied to Y Combinator in 2009. He didn't even get an interview.

Cohen's success in adapting the Y Combinator model to Boulder inspired many other competitors. There are now more than a hundred accelerator programs worldwide, with names such as Seedcamp, DreamIt, 500 Startups, and Tech Wildcatters. TechStars is loosely affiliated with 30 of these programs-;serving in an advisory capacity and allowing the programs to use its application-;and it runs TechStars programs in New York City, Boston, Seattle, San Antonio, and Boulder. "There is a start-up revolution occurring," says Feld. "Every major metro area in the world will eventually be able to support an accelerator."

Two weeks before the TechStars application deadline, 200 or so young people pack into a New York University auditorium, paying $5 each for a class called How to Get Into TechStars NYC. The two-hour forum, led by Cohen and Tisch, focuses largely on the importance of applicants showing that they have more than a good idea. At a time when the cost of launching a Web company is close to zero, a great concept isn't nearly enough. "The best way for us to know that you're going to do stuff is to do stuff," Tisch booms from the front of the room. "Show us evidence of progress."

Most of the companies that are eventually accepted into TechStars get noticed by sending weekly, or even daily, updates about new versions of their products, new customers, and new hires. This process typically happens during the weeks after they have submitted their application, but it can sometimes drag on for years, as has been the case with Frederick Cook, a 26-year-old Virginia native. Five years ago, Cook was taking an engineering class in Virginia Tech's Norris Hall when gunfire broke out in an adjacent room. He jumped from a second-story window as his professor was shot to death in what became known as the Virginia Tech massacre. Cook had planned to work for a defense contractor, but the experience set him down the start-up path. "The shooting changed the way I think about a lot of things in life," Cook says. "And one of them was that I wasn't going to do anything less than my full potential. For me, that was starting a business." He taught himself Web development and, in early 2010, he applied to Y Combinator and TechStars.

His first company was rejected outright by both programs. His second idea earned him an interview with Tisch but not a spot. After that idea was rejected, he scrapped it, came up with a new one in a matter of days, and reapplied. Tisch was impressed, but not enough to let Cook into the program. "We've been tracking him for two years," says Tisch. "This is a kid I desperately wanted to fund. But his company sucked."

Whereas Cook's early efforts were essentially copycat social networking sites, his new company-;a website called Moveline, which helps people compare and hire movers-;has no obvious competitors. Cook has raised $400,000 from investors, recruited a capable co-founder, and, for the first time, Tisch believes, found a business that fits his personality. "One of Fred's challenges is that his energy level is a little low," says Tisch. "But this is a business that a laid-back, thoughtful guy might be able to execute well."

As with most things in life, however, the easiest way to get a leg up with TechStars is to know somebody. That's the case with Diego Zambrano, a Brazilian designer with a Rip Van Winkle beard and a background that would not, at first glance, seem to lend itself to a career as a scrappy technology entrepreneur. He's older than 30, he doesn't know how to write code, and he doesn't have a co-founder.

But none of this matters, because Tisch is fairly certain that Zambrano is going places. Before founding his start-up, Bondsy (tag line: "the social network of things"), Zambrano worked as a creative director at the advertising giant Ogilvy & Mather. "If you become creative director at Ogilvy with that beard, that means you're stupidly talented," says Tisch. More important, as Tisch tells me, Zambrano has already raised funding from a roster of well-known angel investors. "It's not that we're giving certain companies an advantage for no reason," Tisch says. "They've earned the advantage by either knowing us or knowing people who know us. My best friends in the tech world are working on Bondsy." Zambrano, it turns out, hasn't even filled out an application yet.

Tisch seems to instantly regret divulging this last fact. "I don't know if you should write in the article that you can go around the process," he says. "But look, this isn't about applying randomly. It's about building relationships."

Everybody ready?"

Three weeks have passed since the application deadline, and 1,480 applicants have been whittled down to 85 of the strongest, whose stories Tisch is now presenting to the TechStars selection committee. The group includes Cohen, partners from five venture capital funds, and two angel investors. Tisch calls the crowd to attention and begins pitching. "All right, Moveline," says Tisch. "This is a kid we've wanted to back for a while."

Tisch tells them the story of Frederick Cook's escape from the Virginia Tech massacre and then asks them to take his company seriously. He hopes to make a decision as soon as possible. "I promised Fred that we'd give him an answer tomorrow to avoid disappointing him again," Tisch says.

The investors are unmoved.

"Does this create a problem for people that doesn't exist?" one asks, wondering whether there is really a market for online moving services.

"I feel like U-Haul takes all the share in this space."

"Are we just trying to see how many times we can get this kid to apply to TechStars just for shits and giggles?"

Everyone laughs.

"Poor Fred," someone says.

"All right," says Tisch, clearly disappointed. "He's out."

Other casualties of the meeting include a book-publishing start-up whose CEO has an Ivy League M.B.A. and who served as the personal assistant to Ari Emanuel; an augmented-reality company whose 50-ish founder previously started an Inc. 500 company; and the beauty company with the flutey founder.

Bondsy makes it through. The investors, some of whom have met Zambrano, are impressed with his idea, his background, and, of course, his beard. "He's the most interesting person that you'll have in this program," says one.

O ver the course of a week or so, Tisch, who makes the final decisions about whom to accept, spends at least an hour with each of the 30 finalists. In cases where the selection committee has been overwhelmingly positive, he will make an offer on the spot. "We would love to have you as part of TechStars," Tisch tells Zambrano, whose face immediately turns bright red. "I hope you accept."

He does. "Thank you for trusting me," he says.

The next meeting, with a company called Lua Technologies, is testier. "The scariest thing about your team is that you're top-heavy," Tisch says. The company has three co-founders, plus two engineers, a designer, and a marketer. Tisch looks dismissively at the marketing guy and then turns to the company's CEO, Michael DeFranco, and suggests that he might want to fire him. DeFranco stumbles in response but holds his ground. Tisch excuses himself. He returns and makes the team an offer. "We love you guys," he says.

A few hours later, Jay Finch, a 26-year-old applicant from Jersey City, walks into the room. Finch's company, Socstock, which aims to replace traditional small-business loans with crowdsourced funding from customers, is one that Tisch has been eyeing for months. He loves the business model, and he especially likes Finch, because he has a great story. Finch is a former Goldman Sachs analyst who quit his job because he wanted to help neighborhood businesses rather than ridiculously wealthy investors.

Finch is also one of only two black CEOs to make it to the final round. "I genuinely want to fund an African American founder this time," says Tisch. "That's an important box to check. If we don't get there with Jay, we won't get there with anybody else this time." The company got mostly good marks from the selection committee, and Tisch tells me that Finch will probably get an offer.

The only hitch is that a nearly identical company applied to TechStars just before the deadline. The CEO is also a young guy named Jay-;his last name is Lee-;and he has a remarkably similar background. A Harvard Law graduate and former securities lawyer, Lee started Smallknot, which is based in Brooklyn, New York, with a Harvard classmate.

Finch is Tisch's first choice, but the meeting doesn't go particularly well. When Tisch asks Finch to rethink Socstock's repayment process, Finch struggles to offer alternatives. "You're adding all of this weird complexity to the process," Tisch says.

"We're not married to it," Finch says.

Tisch changes the subject. "What do you think of your name?" he asks.

"We like it," says Finch's co-founder, Julissa Arce, a former Goldman Sachs vice president.

Tisch snorts. "That's because you're finance people," he says. "You need to connect emotionally with your users. You have a lot of heart. Take that and put it in your product. It isn't there now."

When Tisch excuses himself, the co-founders look mystified.

"They think too much like bankers," Tisch tells me after Finch and Arce are gone. "I feel like we have to meet with the other team."

I return the following Thursday and find Tisch in a buoyant mood. TechStars has made eight offers so far-;most recently to a Dutch start-up called Karma, which hadn't filled out an application but which (thanks to a recommendation from a prominent VC) seemed too promising to pass up.

The meeting with Smallknot, Socstock's competitor, went well. Tisch's only concern is that Lee hasn't managed to hire a full-time software developer. "You say you're trying to inspire a movement," Tisch's partner, Adam Rothenberg, wrote to Lee in an e-mail. "How come you can't inspire one developer to join your team?"

Meanwhile, Tisch is agonizing about whether to reject Moveline. When he called Cook to tell him about the selection committee's recommendation, Cook defended his idea with aplomb, Tisch tells me. Three days later-;and more than two years after he first applied-;Cook will get his offer.

Thursday night, Smallknot hires a software developer and immediately reports back to Tisch. Tisch sends an e-mail around 1 a.m. Saturday, asking some technical questions about Smallknot's model. Lee wakes up his co-founder, and the two men scramble to craft a response, which they send at 3:43 a.m. They are accepted Sunday evening.

On the other side of the Hudson River, Jay Finch spends the weekend waiting. "When I didn't hear on Sunday, I knew something was up," he says over coffee a few days later. "I was bummed."

When I tell him that a competing company called Smallknot was accepted into the program, he looks crestfallen.

"But they're not really a competitor, right?" he asks. "They don't have the same market as us?"

Yes, I say, they pretty much do. I tell him that his was one of 30 crowdfunding businesses to apply to TechStars this time around-;a fact that I think is going to cheer him up but serves only to make the situation seem more hopeless.

"Well, now those guys have all the resources and connections that TechStars has," Finch says. "That's a problem for us."

He sits there for a minute without saying anything, then slowly begins to regain his composure. He tells me that he will press on, though maybe not in New York. He says he's heard a lot of good things about 500 Startups, a Bay Area incubator, and may apply to that program. He isn't giving up.

Three weeks later, on March 14, the 14 teams-;some 70 hopeful people-;will arrive at the TechStars offices on the sixth floor of the Village Voice building in New York City, taking seats under vinyl banners that bear the names and logos of their newborn companies. They will work nonstop for three months-;the typical approach is to take meetings with mentors during the day and use nights and weekends to write code. They won't sleep more than a few hours a night.

And, despite Tisch's best efforts to pick winners, they won't all succeed. "It's hard," he says. "You're going off so little data, and these companies are so early that you can't feel utterly confident in the bets you make. Once they start the program, that's when you get to know how well you picked them. That's when the pressure starts."